Kellanova
predictive scoring of creative effectiveness before delivery
Kellanova ran a 12-month study with Vidmob and MMA Global on 443 creative assets (10 brands, Meta paid social) showing predictive scoring of the 3-second VTR at 83% accuracy, performance multiplied by 2.16, and Profit ROI up 11% on scored assets.
Key points
- Predictive scoring of creative effectiveness before delivery, on 443 assets and 10 brands.
- Vidmob Aperture platform with machine learning and Google Gemini.
- 3-second VTR predicted at 83%, performance 2.16x, Profit ROI +11%.
- Evidence B, status confirmed.
Objective
Understand which creative elements make an ad perform, predict effectiveness before delivery, guide creative production, and rebuild the agency relationship on a quantitative rather than subjective metric.
The deployment
Kellanova, owner of Pringles, Pop-Tarts, Cheez-It, and Rice Krispies Treats, ran with Vidmob and the association MMA Global a twelve-month study on the performance of 443 creative assets, spread across ten brands, delivered in the United States on Meta ad units. Vidmob's Aperture platform relies on machine learning, historical data, and LLMs including Google Gemini. The analysis established 19 cross-cutting scoring criteria and 11 per-category criteria, common to assets reaching a view-through rate of at least three seconds, then built two predictive models (one for savory, one for sweet). Kellanova reports that the predictive scoring forecasts the three-second VTR with 83 percent accuracy, improves performance by 2.16x, and contributes to an 11 percent rise in Profit ROI on scored assets. The company uses these results to tighten its creative production and to move from a subjective metric to a quantitative one in its agency relationships. The framework is first applied to paid social, with an extension planned to TikTok, Pinterest, and Reddit.
Results Proof B
Quantified study published jointly by Vidmob, Kellanova, and MMA Global (release via Businesswire), corroborated by trade press (Digiday, Adweek) naming Kellanova and its figures. No isolated financial results, so B.
How it works
Documented architectureThe stack in detail
- plateforme Vidmob Aperture creative analytics platform: tagging of creative elements and predictive impact scoring before delivery
- llm Google Gemini LLM used by Aperture to analyze the creative assets
- llm Modeles predictifs par categorie (salle, sucre) two ML models trained on the creative criteria correlated with the 3-second VTR, 83% accuracy
- infra Donnees de performance Meta 3-second VTR and performance of Meta ad units, the training and validation signal
- integrateur MMA Global industry association that framed the twelve-month study with Kellanova and Vidmob
How it runs, concretely
For ops teams-
1Collection and tagging AI / Vidmob
Analysis of 443 assets across ten brands, extraction of 19 cross-cutting criteria and 11 per category.
-
2Model building AI / data team
Two predictive models (savory, sweet) trained on the criteria correlated with the 3-second VTR.
-
3Scoring before delivery marketing / AI
Each new asset gets a predictive impact score before going into media.
-
4Production and agency relationship marketing
The insights guide creative production and serve as a quantitative metric in agency contracts.
The three-second view-through rate on Meta, the reference benchmark. Without enough historical performance data, the predictive models lose accuracy.
How your customers perceive this type of use
Sourced studiesC'est la famille la moins acceptee : 68% des Americains jugent inacceptable un score financier personnel calcule par algorithme et 67% l'analyse video automatisee d'entretiens d'embauche (Pew Research, 2018). La demande d'explication et de recours est massive : 83% veulent savoir quelles donnees l'IA utilise et 91% veulent pouvoir corriger des donnees erronees (Consumer Reports, 2024). A l'echelle mondiale, seuls 46% se disent prets a faire confiance aux systemes d'IA et 70% jugent une regulation necessaire (KPMG / Universite de Melbourne, 2025).
Acceptance conditions
- Transparence sur les donnees utilisees : 83% des Americains la reclament (Consumer Reports 2024)
- Droit de correction des donnees erronees : 91% le demandent (Consumer Reports 2024)
- Explication de la logique de decision : 44% des consommateurs sont plus enclins a utiliser un agent IA si sa logique est clairement expliquee (Salesforce 2024)
- L'acceptabilite depend du contexte de la decision : 50% des Americains jugent equitable un score de risque criminel pour la liberation conditionnelle, contre 32% pour un score financier applique aux consommateurs (Pew Research 2018)
Red lines
- La decision opaque et sans recours sur l'emploi, le credit ou le logement : 45% tres mal a l'aise pour l'embauche, 39% pour le pret, 39% pour le logement (Consumer Reports 2024)
- Le scoring des personnes a partir de donnees comportementales : 68% le jugent inacceptable pour les offres financieres (Pew Research 2018)
Sources: Pew Research Center 2018 · Consumer Reports 2024 · KPMG / Universite de Melbourne 2025 · Salesforce 2024
How to replicate
Inference, not sourcedData prerequisites
- history of tagged creative assets
- media performance data per asset (VTR, views)
- enough volume to train per-category models
Org prerequisites
- alignment of media / creative / agencies on a common metric
- governance of creative production driven by the score
Possible stack
- creative analytics platform (Vidmob Aperture type)
- LLM for asset analysis
- connectors to media platforms (Meta, TikTok)
The plan, step by step
- Step 1Gather the history of assets and the media performance per asset (aim for several hundred assets)Deliverable: Consolidated asset-performance dataset
- Step 2Tag the creative elements against a grid of cross-cutting and per-category criteriaDeliverable: Tagged assets with a documented grid
- Step 3Correlate criteria and 3-second VTR, build a predictive model per categoryDeliverable: Model with accuracy measured on a holdout sample
- Step 4Score new assets before going into media and guide productionDeliverable: Pre-delivery scoring process integrated into the creative workflow
- Step 5Make the score a contractual metric with agencies and extend to other platformsDeliverable: Quantitative metric in the briefs and the agency contracts
First step: Tag a history of paid social assets and correlate the creative elements with the 3-second VTR.
The same pattern, proven in other industries
Bayer Consumer Health
machine learning demand forecasting on search trends to drive media
Rakuten
prediction of future buyers for ad targeting
Rovio
LTV prediction to steer acquisition (pLTV)
Sources
- S1 Behind Kellanova's AI-powered push to improve creative and alter agency fees Established press archive pending
- S2 Vidmob, Kellanova, and MMA Global Collaborate on a Study Proving the Power of Predictive Impact Scoring Interested party archive pending
- S3 Vidmob, Kellanova, and MMA Global Release Study Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.